This invention generally relates to analyzing images, and more specifically to identifying regions of an image free of objects prior to presentation to a user.
A digital magazine server provides digital content to its users as a digital magazine including various pages that each include content items arranged in positions relative to each other. Frequently, digital content provided to digital magazine server users includes images. To increase the number of content items presented in a page of the digital magazine, the digital magazine server often includes text providing additional information about an image in the image itself. However, conventional methods of including text in an image often result in the text being illegible or obscuring objects in the image likely to be of interest to a user.
Additionally, the digital magazine server may crop images to reduce the size of images presented to users by removing portions of the image. However, conventional methods of cropping images often crop objects in the image, reducing the information conveyed to a user by the image. For example conventional image cropping methods merely remove portions of an image based on image size, regardless of the content in different portions of the image.